|Publication number||US5212742 A|
|Application number||US 07/705,284|
|Publication date||18 May 1993|
|Filing date||24 May 1991|
|Priority date||24 May 1991|
|Also published as||US5461679|
|Publication number||07705284, 705284, US 5212742 A, US 5212742A, US-A-5212742, US5212742 A, US5212742A|
|Inventors||James O. Normile, Chia L. Yeh, Daniel W. Wright, Ke-Chiang Chu|
|Original Assignee||Apple Computer, Inc.|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (5), Non-Patent Citations (21), Referenced by (369), Classifications (53), Legal Events (6)|
|External Links: USPTO, USPTO Assignment, Espacenet|
1. Field of the Invention
This invention relates to the field of video imaging systems. More specifically, this invention relates to an improved method and apparatus for video encoding/decoding.
2. Description of the Related Art
Due to the storage requirements, recent demands for full motion video in such applications as video mail, video telephony, video teleconferencing, image database browsing, multimedia, and other applications have required that standards be introduced for video compression. One image of 35 mm slide quality resolution requires 50 megabytes of data to be represented in a computer system (this number is arrived at by multiplying the horizontal by the vertical resolution by the number of bits to represent the full color range or 4096×4096×8×3 [R+G+B] 18=50,331,648 bytes). One frame of digitized NTSC (National Television Standards Committee) quality video comprising 720×480 pixels requires approximately one half megabyte of digital data to represent the image (720×480×1.5 bytes per pixel). In an NTSC system which operates at approximately 30 frames per second, digitized NTSC-quality video will therefore generate approximately 15.552 megabytes of data per second. Without compression, assuming a storage capability of one gigabyte with a two megabytes per second access rate, it is possible to:
a. store 65 seconds of live video on the disk and to play it back at 3 frames per second;
b. store 21 high quality still images taking 24 seconds to store or retrieve one such image.
Assuming that a fiber distributed data interface (FDDI) is available with a bandwidth of 200 megabits per second, 1.5 channels of live video can be accommodated, or 35 mm quality still images can be transmitted at the rate of one every two seconds. With currently available technology in CD-ROM, a likely distribution medium for products containing video, the current transfer rate is approximately 0.18 megabytes per second. 0.37 megabytes per second may be attained with CD-ROM in the near future.
For illustration, take the variable parameters to be the horizontal and vertical resolution and frame rate, and assume that 24 bits are used to represent each pixel. Let D represent the horizontal or vertical dimension and assume an aspect ratio of 4:3. The data rate in megabytes per second as a function of frame rate and image size is:
______________________________________Image Size Frame Rate per secondD 5 10 15 20 25 30______________________________________ 64 0.04 0.08 0.12 0.16 0.20 0.24128 0.16 0.33 0.49 0.65 0.82 0.98256 0.65 1.31 1.96 2.62 3.27 3.93512 2.62 5.24 7.86 10.48 13.10 15.72______________________________________
or formulated in a slightly different way, the number of minutes of storage on a 600 megabyte disk is:
______________________________________Image Size Frame Rate per secondD 5 10 15 20 25 30______________________________________ 64 244.20 122.10 81.40 61.06 48.84 40.70128 61.05 30.52 20.35 12.25 12.21 10.17256 15.26 7.63 5.08 3.81 3.05 2.54512 3.81 1.90 1.27 0.95 0.76 0.63______________________________________
It is obvious from data rate and storage considerations that data compaction is required in order for full motion video to be attained.
In light of these storage and rate problems, some form of video compression is required in order to reduce the amount of storage and increase the throughput required to display full-motion video in a quality closely approximating NTSC. Photographic and, to an even greater degree, moving images generally portray information which contains much repetition, smooth motion, and redundant information. Stated in an equivalent way, areas of an image are often correlated with each other, as are sequences of images over time. Keeping these facts in mind, several techniques as have been established which eliminate redundancy in video imaging in order to compress these images to a more manageable size which requires less storage, and may be displayed at a fairly high rate. Some simple compression techniques include:
1. Horizontal and Vertical Subsampling: Sampling only a limited number of pixels horizontally or vertically across an image. The required reduction in resolution provides for poor quality images.
2. Reduction in Number of Bits Per Pixel: The technique including the use of a Color Look Up Table is currently used successfully to reduce from 24 to 8 bits per pixel. A reduction of approximately 3-1 is the useful limit of this method.
3. Block Truncation Coding and Color Cell Methods: The block truncation coding (BTC) was developed by Bob Mitchell in the early 1980's targeted at low compression rate and high quality applications (Robert Mitchell, et al., Image Compression Using Block Truncation Coding, IEEE Trans., Comm., pp. 1335-1342, Vol. Com-27, No. 9, Sep. 1979). In this scheme, the first order statistics (mean) and the second order statistics (variance) of each pixel block is extracted and transmitted. The image is reconstructed using these two quantities. An 8-1 compression ratio with 4×4 block sizes was demonstrated in (Graham Campbell, Two Bit/Pixel Full Color Encoding, pp. 215-223, Proceedings of SIGGRAPH '86, Vol. 20, No. 4, Aug. 1986).
4. Vector Quantization (VQ): A simple VQ maps discrete k-dimensional vectors into a digital sequence for transmission or storage. Each vector (a block of 4×4 or 3×3 pixels) is compared to a number of templates in the code book, and the index of the best matched template is transmitted to the receiver. The receiver uses the index for table look-up to reconstruct the image. A simple VQ could provide about 20-1 compression with good quality. A more complex VQ scheme has been demonstrated to provide similar quality to the CCITT (International Consultative Committee for Telephony & Telegraphy) DCT (Discrete Cosine Transformation) scheme recommendation H.261 (T. Murakami, Scene Adaptive Vector Quantization for Image Coding, Globecom, 1988).
5. Predictive Techniques: The assumption on which this family of methods relies is that adjacent pixels are correlated. As a consequence, data reduction can be accomplished by predicting pixel values based on their neighbors. The difference between the predicted and the actual pixel value is then encoded. An extensive body of work exists on this technique and variations on it (O'Neil, J. B., Predictive Quantization Systems for Transmission of TV Signals, Bell System Technical Journal, pp. 689-721, May/June 1966).
The compression ratio to be expected from each of these simple methods is between four and eight to one.
More complex techniques for video compression are also known in the art. It is possible to achieve data compression of between four and eight to one by using some of the simpler techniques as mentioned above. To achieve comparable quality, at compression ratios from twenty to forty to one, involves a superlinear increase in complexity. In this case, it is no longer appropriate to consider the compression process as a simple one-step procedure.
In general, lossless compression techniques attempt to whiten or decorrelate a source signal. Intuitively, this makes sense in that a decorrelated signal cannot be compressed further or represented more compactly. For compression ratios of greater than twenty to one, a lossy element must be introduced somewhere into the process. This is usually done through a temporal or spatial resolution reduction used in conjunction with a quantization process. The quantization may be either vector or scalar. The quantizer should be positioned so that a graceful degradation of perceived quality with an increasing compression ratio results.
Many of the succeeding methods are complex, but may be broken into a series of simpler steps. The compression process can be viewed as a number of linear transformations followed by quantization. The quantization is in turn followed by a lossless encoding process. The transformations applied to the image are designed to reduce redundancy in a representational, spatial and temporal sense. Each transformation is described individually.
Although the RGB representation of digitized images is common and useful, it is not particularly efficient. Each one of the red, green and blue constituents is potentially of full bandwidth, although much of the information is correlated from plane to plane. The first step in the compression process is to decorrelate the three components. If an exact transformation were used, it would be image dependent, and as a consequence computationally expensive. A useful approximation which does not depend on image statistics is the following: ##EQU1## In the case of NTSC images, the resulting U and V (chrominance components containing color information) components are of lower bandwidth than the Y (luminance component containing the monochrome information). In general, the U and V components are of less perceptual importance than the Y component. The next stage in the compression process usually consists of subsampling U and V horizontally and vertically by a factor of two or four. This is done by low pass filtering followed by decimation. At this point in the process, much of the interplane redundancy has been removed, and a data reduction by a factor of two has been achieved.
Reduction of temporal redundancy may be achieved simply by taking the difference between successive frames. In the case of no motion and no scene change, this frame difference will be zero. The situation is more complex when there is interframe motion. In this case, some reregistration of the portions of the image which have moved is required prior to frame differencing. This is done by estimating how far pixels have moved between frames. Allowing for this movement, corresponding pixels are again subtracted (Ronald Plompen, et al., Motion Video Coding in CCITT SG XV--The Video Source Coding, pp. 997-1004, IEEE Global Telecommunications Conference, December 1988). The motion vectors are coded and transmitted with the compressed bit stream. These vectors are again used in the decoding process to reconstruct the images. The distinction between frame differencing and differencing using motion estimation may be expressed as follows. In the case of simple differencing, the error between frames is calculated as:
using motion estimation error may be written as:
where x and y are the calculated displacements in the x and y directions respectively.
In most images, adjacent pixels are correlated. Reducing spatial redundancy involves removing this correlation. The removal is achieved using a linear transformation on the spatial data. In the ideal case, this transform should depend on image statistics. Such a transform does exist and is known as the Hotelling or Karhounen Loueve (KL) transform (N. S. Jayant, Peter Noll, Digital Coding of Waveforms, Prentice Hall, Signal Processing Series, p. 58). As it is computationally expensive and does not have a fast algorithmic implementation, it is used only as a reference to evaluate other transforms. A variety of other transforms have been applied to the problem, including: Fourier, Walsh, Slant, Haddamard (Arun, N. Netravali, Barry G. Haskell, Digital Pictures Representation and Compression, Plenum Press). The cosine transform provides the best performance (in the sense of being close to the KL transform). The discrete cosine transform is defined in the following way: ##EQU2## where x(m,n) is an N×N field (Blocksize), k, l, m, n all range from 0 to N-1, and ##EQU3## α(j)=1; j≠0. A range of DCT "block sizes" have been investigated for image compression (A. Netravali, et al., Picture Coding: A Review, Proc. IEEE, pp. 366-406, March 1980), and standards bodies have decided, apparently in an arbitrary fashion, that 8×8 blocks are "best." Adaptive block size have also been considered, with the adaptation driven by image activity in the area to be transformed (see Chen, C.T., "Adaptive Transform Coding Via Quad-Tree-Based Variable Block Size," Proceedings of ICASSP '89, pp. 1854-1857). In summary, a combination of the above techniques, as applied to a raw video image would be performed as follows:
1. Digitizing the image;
2. transform RGB to YUV;
3. remove temporal redundancy (through frame differencing and motion compensation;
4. remove spatial redundancy (through a discrete cosine transfer); and
5. entropy encode the data (using Huffman coding).
This process yields the maximum compression possible using prior state of the art techniques.
Three examples of state of the art compression methods using some of these techniques are known as: CCITT H.261 (the International Consultive Committee for Telephony and Telegraphy); JPEG (Joint Photographers Experts Group); and MPEG (the Motion Picture Experts Group). The JPEG standard was developed for encoding photographic still images and sets forth a basic technique for encoding video image data. The technique converts 8×8 pixel blocks of the source image using a discrete cosine transformation (DCT) function, with each block of pixels being represented in YUV source format (representing luminance and chrominance information for the block). Threshold blocks of DCT coefficients using psychovisual thresholding matrices are then used to quantize the results of the 8×8 DCT macroblocks of the source image. Finally, each of the blocks is entropy encoded. The decoding process reverses these steps.
The CCITT H.261 standard was developed for use in video teleconferencing and video telephony applications. It can operate at 64 kilobits (Kbits) per second to 1.92 megabits (Mbits) per second, and can operate upon images between 525 and 625 lines based upon a common intermediate format (CIF). It is performed using a method as shown in FIG. 1.
The CCITT encoder 100 consists of a DCT, a zig-zag scanned quantization, and Huffman coding. DCT 101, quantizer 102, and variable length coding 103 blocks perform the coding function. Finally, multiplexer 104 combines the Huffman code from the variable length coding block 103, motion vector data from motion estimation block 105, quantizer data from quantizer block 102. Intra/Inter type information from intra/inter block 106 and performs formatting and serializing, video synchronization and block addressing. Frame memory 107 is used to determine differences from the previous frame and the current frame using motion estimation block 105. CCITT encoder 100 further comprises inverse quantizer 108 and inverse DCT function 109 to provide frame difference information. Lastly, information multiplexed by 104 is passed to rate control 111 and buffer 112 for output as compressed bit stream 120.
The CCITT decoder is shown in FIG. 2 as 200. Demultiplexing block 201 takes the encoded bit stream 210, identifies its constituents and routes them to the relevant parts of decoder 200. The main function of variable length decoding block 202 and inverse quantizer 203 block is to reconstruct the DCT coefficients from their Huffman encoded values, rescale these values and pass these on to inverse DCT block 204. Inverse DCT block 204 takes coefficients in blocks of 8×8 and reconstitutes the underlying spatial video information. If the macro block is intra-coded, no motion estimation is invoked. If it is inter-coded, the output is a difference between the information in this frame and the motion-compensated information in the last frame. A motion vector transmitted from demultiplexer 201 via "side information" signal 208 determines the best block to be used for reconstruction from the last frame. Motion compensation is performed by 206 from information of the current image in frame buffer 207. This is fed back into the decoded stream 205 and then as decoded output information 220 in CIF format. The Y and UV components share the same motion vector information. A detailed description of the CCITT H.261 standard is described in document No. 584, published on Nov. 10, 1989 by the Specialists Group on Coding For Visual Telephony, entitled Draft Revision of Recommendation H.261 published by the CCITT SG XV Working Party XV/1 (1989).
The MPEG standard is the most recent compression specification to use transport methods which describe motion video. Though not fully finalized, the MPEG specification's goal is to obtain VHS quality on reconstruction of the images with a bit rate of approximately 1.15 megabits per second for video. This yields a total compression ratio of about 40-1. The distinguishing feature of MPEG from JPEG and CCITT H.261 is that MPEG provides a higher quality image than CCITT H.261, like JPEG but allows motion. This is in contrast to JPEG, which only provides still-frame imagery and no audio. In addition, MPEG adds the additional feature of synchronized sound along with the encoded video data although it has not been finalized. A detailed description of MPEG may be found in the document entitled MPEG Video Simulation Model 3 (SM3)--Draft No. 1 published by the International Organization for Standardization ISO-IEC/JTC1/SC2/WG8, Coded Representation of Picture and Audio Information ISO-IEC/JTC1/SC2/WG8 N MPEG 90/, published by A. Koster of PTT Research.
Some of the relative advantages and disadvantages of the various coding algorithms are set forth as follows. JPEG provides no description of motion video at all. MPEG, although a full featured standard (it provides both forward motion, backwards motion, and still frame), is still under development and undergoing revision. CCITT H.261, because it was developed for teleconferencing and video telephony, it provides a moving source but has no provisions for viewing the motion picture images in a reverse direction, or provides any means for still frame viewing. Therefore, a system is required which is fairly mature, such as the CCITT H.261 standard, but yet provides all the capabilities (including reverse play and still frame) of a full-featured compression system, such as the MPEG standard.
CCITT H.261 uses a scheme such as that shown in FIG. 3 in order to provide for full-motion video. FIG. 3 shows a series of frames which represents a particular section of moving video. 301 and 302 contain full scene information for the image at the beginning of a series of frames. 301 and 302 are known as "intra" frames or keyframes which are used in CCITT H.261. Each intra frame 301 or 302 contains a full scene description of the frame at the times they appear. Although compressed, intra frames 301 and 302 contain substantial information. Each of the intervening frames between two intra frames 301 and 302 are known as "inter" frames 303, 304, and 305. Each inter frame such as 303-305 contains information which should be added to the preceding frame. For example, inter frame 303 only contains information which has moved since intra frame 301. Therefore, the information represented in frames 303-305 may be substantially less than that contained in frames 301 and 302 because the inter frames contain only motion data, and not complete scene information for the entire frame. This provides a fairly high compression ratio for intervening inter frames 303-305. CCITT H.261 as represented in FIG. 3 is incapable of providing reverse motion video because a "key" frame, such as intra frame 301 only establishes information for inter frames 303-305 which follow intra frame 301 in time. In other words, 303-305 only contain information which has moved from intra frame 301, not motion information from intra frame 302. An attempt to play such a sequence of frames in reverse will generate substantial distortion of the moving image.
Because a decompression rate of approximately 30 frames per second (FPS) is required for real-time moving video, the processor performing the decoding process must have a fairly high bandwidth and be able to handle all the necessary matrix-matrix multiply operations required by the decoding process in a short period of time. To date, no single device possesses the necessary computing power to decompress an incoming compressed bit stream at the necessary rate to make data available for NTSC quality video at the 30 frame per second rate.
One of the objects of the present invention is provide an architecture and method which has sufficient computing power to allow compressed moving video images to be decompressed and displayed in real time.
This and other objects of the present invention are provided for by an apparatus for processing video data for compression/decompression in real-time which comprises a plurality of compute modules, in a preferred embodiment, for a total of four compute modules coupled in parallel. Each of the compute modules has a processor, dual port memory, scratch-pad memory, and an arbitration mechanism. In a preferred embodiment, the processor is a digital signal processor, and the device comprises 16 kilobytes of dualport dynamic random access memory and 64 kilobytes of local "scratch pad" dynamic random access memory. A first bus couples the compute modules and a host processor. In a preferred embodiment, the host processor is coupled to a complete computer system comprising a display, memory, and other peripherals, and the first bus is known as a control bus which operates at a relatively low speed. Lastly, the device comprises a shared memory which is coupled to the host processor and to the compute modules with a second bus. This second bus is known in a preferred embodiment as a "video" bus and operates at substantially higher speeds than the "control" bus. The shared memory, in a preferred embodiment, comprises two megabytes of static random access memory to allow access by both the compute modules via the video bus, and the host processor, the access being controlled by the arbitration mechanism in a first in first out order (FIFO), which arbitration mechanism is a gate array, or other discrete circuitry in a preferred embodiment. In an alternative embodiment, there is a frame buffer coupled to the "video" bus, and a display means coupled to the frame buffer, which acts as the display of the system, instead of that normally connected to the host. This provides increased performance of the system as a whole, especially in video decompression tasks.
These and other objects of the present invention are provided for by a method in a computer system for partitioning an image for processing by a parallel computing system. The parallel computing system comprises N computing units. First, the total length of an image is determined, and is divided by N. The dividend is then stored in a first value, the first value, in a preferred embodiment, the width of the image to be assigned to each parallel computing unit. A first region of the image is assigned to a first computing unit, the first region starting at a first position, and ending at the first position offset by the first value. Therefore, a portion of the image is assigned, in a preferred embodiment which is the full image in width, and H/N wherein H is the length or height of the image, and N is the total number of processors. Each of the N processors are assigned corresponding sections of the image according to their position relative to the first processor, each having a section which is the full width of the image, and which is H/N in length. Height and width information is represented, in the preferred embodiment, in blocks containing luminance and chrominance information.
The present invention is illustrated by way of example and not limitation in the figures of the accompanying in which like references indicate like elements and in which:
FIG. 1 shows a prior art video encoder.
FIG. 2 shows a prior art video decoder.
FIG. 3 shows a prior art scheme for representing motion video.
FIG. 4 shows the architecture of the video processor of the preferred embodiment.
FIG. 5 shows one compute module used in the preferred embodiment of the present invention.
FIG. 6 shows the partitioning of an image for processing by each of the computing modules of the preferred embodiment.
FIG. 7a shows the preferred embodiment's method of encoding motion video.
FIG. 7b is a detailed representation of one frame in the preferred embodiment.
FIG. 8a shows the improved CCITT encoder used in the preferred embodiment.
FIG. 8b is a detailed representation of the scene change detector of the preferred embodiment.
FIG. 9 shows the enhanced CCITT decoder used in the preferred embodiment.
The present invention relates to a method and apparatus for video encoding/decoding. In the following description, for the purposes of explanation, specific values, signals, coding formats, circuitry, and input formats are set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, to one skilled in the art, the present invention may be practiced without these specific details. In other instances, well known circuits and devices are shown in block diagram form in order to not unnecessarily obscure the present invention.
A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by any one of the patent disclosure, as it appears in the Patent and Trademark Office patent files of records, but otherwise reserves all copyright rights whatsoever.
Referring to FIG. 4, an architecture of a parallel processing system which is used for compression/decompression of moving video images in the preferred embodiment is shown as 400. The architecture of the preferred embodiment provides a parallel coupling of multiple video processing modules such as 401-404 which has the necessary bandwidth to decompress video images at the frame rates required by motion video (for instance, 30 frames per second). Modules 401-404 are coupled to a computer system bus 425 via control bus 412 in the preferred embodiment. Also, coupled to system bus 425 is display controller 426, which is coupled to frame buffer 427. Frame buffer 427 in turn is coupled to display 426 for displaying information. In the preferred embodiment, information is placed onto bus 425 by modules 401-404 or host processor 410, and read in by display controller 426 for placing into frame buffer 427 and display on 428. Although host processor 410, in the preferred embodiment, is typically the bus master of system bus 425, at certain times display controller 426 assumes control of system bus 425. Display controller 426 can increase the typical throughput on bus 425, to allow uncompressed data to be received from modules 401-404 and placed into frame buffer 427 in the required time. In the preferred embodiment, display controller 426 is an AMD 29000 RISC processor manufactured by Advanced Micro Devices of Sunnyvale, Calif. Host processor 410 is one of the 68000 family of microprocessors such as the 68030 or 68020 manufactured by Motorola, Inc. of Schaumburg, Ill. System 400 shown in FIG. 4 is implemented in a computer system such as one of the Macintosh® family of personal computers for example the Macintosh® II, manufactured by Apple® Computers, Inc. of Cupertino, Calif. (Apple® and Macintosh® are registered trademarks of Apple Computer, Inc. of Cupertino, Calif.). System bus 425 is a standard computer system bus capable of operating at 10 MHz which is well-known to those skilled in the art and is capable of transferring data at a maximum rate of approximately 18 Mbytes per second.
System 400 of the preferred embodiment provides a shared memory space 405 which comprises two megabytes (approximately 512,000 entries of 32 bit longwords) of static random access memory (SRAM). Shared memory 405 is coupled to bus 425 via signal lines 411 and is further coupled to high speed "video" bus 420. Video bus 420 may transfer data at a maximum rate of approximately 80 megabytes per second and is clocked at 20 MHz. Control bus 411 transfers data at a maximum rate of ten megabytes per second and is clocked at 10 MHz. Shared memory 405 is separated into two distinct areas. Two banks of equal size are provided in 405 for high data rate ping-ponging between host information and information put into the shared memory by computing modules 401-404.
In addition, memory 405 comprises a small mail box area for communication and task synchronization between host 410 and computing modules 401-404. The mailbox area may vary in length depending on the job size and ranges from one longword (32 bits) to the entire length of memory 405. Although only four parallel computing units 401-404 are set forth in the preferred embodiment, more or less than four parallel processing units may be used along with the corresponding increase or decrease in computing power associated with the addition or the loss of each computing module.
Computing modules 401, 402, 403, and 404 are also connected to a "low" speed control bus 412 which is coupled to each of the computing modules and bus 425 for communication between the computing modules and task synchronization with host processor 410. In addition, in an alternative embodiment, a separate frame buffer 430 may be directly coupled to the video bus 420, which frame buffer may then be then be coupled to a display such as 440 shown in FIG. 4. Although, in the preferred embodiment, the display capabilities of display controller 426, frame buffer 427 and display 428 would normally be used for displaying information processed by modules 401-404, in an alternative embodiment, 430 and 440 may be used by each module 401-404 depositing uncompressed data directly into frame buffer 430 for display. A more detailed representation of one of the computing modules such as 401-404 is shown in FIG. 5.
FIG. 5 is a block diagram of one compute module 401 of the system shown in FIG. 4. Because the remaining compute modules are essentially identical (except for codes embedded in arbitration mechanism 502 to provide their unique identification), only compute module 401 is described in FIG. 5. The structure of module 401 has equal application to each of the remaining compute modules such as 402-404 shown in FIG. 4. Each compute module such as 401 comprises 16 kilobytes (approximately 4,000 32-bit longwords) of dual-port random access memory 504 as shown in FIG. 5. Dual-port memory 504 is coupled to control bus 412 for communication between computing module 401 and devices over bus 425 such as host 410 and display controller 426. Dual-port memory 504 is further coupled to internal bus 530 which is coupled to digital signal processor 501 of compute module 401. Digital signal processor (DSP) 501 provides all processing of data and communication between the various portions of compute module 401. DSP 502 provides the encoding and decoding functions of video data which were discussed with reference to FIGS. 1 and 2. DSP 501, as discussed above, is coupled to dual-port memory 504 through internal bus 530. 530 is a full 32-bit wide data bus and provides communication to host 410 through dual-port memory 504. DSP 501 is further coupled to an internal high speed bus 520 for communication with arbitration mechanism 502 and local memory 503. In the preferred embodiment, 501 is a TMS 320C30 digital signal processor manufactured by Texas Instruments (TI) of Dallas, Tex. Digital signal processor 501 is coupled to local memory 504 via local bus 520. Local memory 503 provides a working or scratch pad memory for the various processing operations performed by DSP 501. Local memory 503 comprises 64 kilobytes of dynamic random access memory (or approximately 16,000 32-bit longwords).
In addition, DSP 501 is coupled to arbitration mechanism 502 through local bus 520. Arbitration mechanism 502 provides circuitry to grant compute module 401 access to shared memory 405 as shown in FIG. 4 over bus 420. Each of the computing modules 401-404 and host processor 410 has a unique identification number in arbitration mechanism 502 shown in FIG. 5. Requests and grants of access to shared memory 405 are performed as follows. The arbitration provided by this logic, allows access to the shared memory and host processor in first-in-first-out (FIFO) order according to each device's bus request (BRQ) number. Host 410 has a bus request number of BRQ0. Access is obtained when bus acknowledge (BACK) is issued. However, if simultaneous requests are made after reset by compute modules 401-405 and host 410, then host 410 gets priority, and obtains access to shared memory bus 405. Compute module 401 is second in priority (BRQ1), module 402 is third (BRQ2), module 403 is fourth (BRQ3), and module 404 is last (BRQ4). In an alternative embodiment, host processor 410 may be given interrupt capability wherein operations being performed by computing modules 401-404 are preempted by host processor 410. In a preferred embodiment, arbitration mechanism 502 is implemented in a gate-array device. The code for this arbitration scheme is set forth in appendix I.
In the preferred embodiment, decompressed data is read by display controller 426 from parallel processing system 400 and transmits the data as uncompressed image data to frame buffer 427. Essentially, display controller 426 when accessed to dual port memory 504 in each of the computing modules 401-404 is available, display controller 426 assumes control of system bus 425 thus becoming the bus master. When display controller 426 becomes the bus master, uncompressed data is read from modules 401-404 and each of their respective dual port memories 504, and that information is placed onto bus 425. Once display controller 426 assumes control of bus 425, data may be transferred on bus 425 at the maximum possible rate allowed by the bus. In the preferred embodiment, the rate is approximately 18 megabytes per second. At that time, host 410 does not participate in the data transfer at all. Once the information is received over bus 425 from modules 401-404, that information is passed to frame buffer 427. Thus, the frame becomes available for display 428 at screen refresh time. The enhanced capabilities of display controller 426 allows the uncompressed data to be available from each of the modules 401-404 at the required 30 fps rate to display 428. Once the entire frame has been made available in frame buffer 427, display controller 426 relinquishes control of bus 425 to host 410. Then, the next cycle of retrieving compressed data and transmitting it to computing modules 401-404 may be performed. In the preferred embodiment, host 410 is required for reading compressed data from disk or memory, transferring the data to modules 401-404, and for servicing user requests.
In order for computing modules 401-404 to decode compressed input image data, the task must be split into component parts to allow each processor to independently process the data. In one scheme, for moving video data where there are N processors, every Nth frame may be assigned to a separate processor. This scheme is not desirable because data from a preceding frame is required to compute the new frame in most decompression algorithms, such as CCITT H.261 and MPEG. Therefore, the preferred embodiment uses a scheme such as that set forth in FIG. 6. The task of splitting this decompression problem into components for each of the computing nodes 401-404 is shown and discussed with reference to FIG. 6. Task synchronization between host 410 and modules 401-404 is discussed in more detail below. An image such as 601 which is displayed at time t is divided into a number of horizontal "stripes" 602, 603, 604, and 605 each of which is assigned to a separate parallel computing node such as 401-404. If 601 is viewed as one complete image or frame, then module 401 will be assigned stripe 602 for processing. Module 402 receives stripe 603, module 403 receives 604, and module 404 receives stripe 605 of the frame for processing. The stripe width is the full screen width of frame 601 shown in FIG. 6, and the stripe length is represented as h wherein h=H/N. H is the total length of the frame and N is the number of parallel computing nodes in video imaging system 400. In addition, each parallel computing node is assigned an image overlap area such as 606 for stripe 602 that overlaps with the next processor stripe such as 603. This allows certain areas of stripes to be shared between computing modules, so that a vertically moving area such as 607 may already be in the local memory of the next processor if the area transitions from one stripe to another. For instance, as shown in FIG. 6, at time t, a moving area 607 in frame 601 may move at time t+1 as shown in frame 611 to a second position which is now in stripe 603. This stripe is handled by the next computing module 402 and already resides in that node's local memory because 607 resided in overlap area 608 at time t. Area 608 was the overlap area for stripe 603, which was assigned to computing module 402. Therefore, at time t+1 for image 611 shown in FIG. 6, computing module 402 will have image data for 607 available in its local memory 503 due to the presence of 607 in overlap area 608 at time t of image 601.
In addition to this frame partitioning scheme, each processor has access to the remainder of the image through shared memory 405 discussed with reference to FIG. 4. The allocation scheme discussed with reference to FIG. 6 will generally provide immediate access to data in vertically overlapping areas through local memory 503 of each computing module. If information is required by the computing module that is outside its stripe, (i.e. a moving area has vertically traversed outside the "overlap" areas, and therefore is not in local memory) then the information may be retrieved from shared memory 405 shown in FIG. 4. This, of course, is achieved at a higher performance penalty because the arbitration mechanism must allow access and delays may occur over bus 420 while accessing shared memory 405. Using this partitioning scheme, each computing module performs inverse DCT, motion compensation and YUV (luminance/chrominance) to RGB functions independently of the other processing modules.
Task synchronization between host 410 and modules 401-404 is now discussed. The preferred embodiment employs a scheme wherein one module such as 401, 402, 403, or 404 is the "master". Once per frame, the master will request host 410 to place in dual port RAM 405 one complete frame of compressed data. The master then decodes the compressed data into "jobs" for individual slave modules, and posts them into shared memory 405.
The 512K longwords in shared memory 405 are logically divided into two areas called "job banks." In each bank resides a stack of "jobs" for individual slave modules. At any given time, the stack of jobs in one bank is being built up by newly decoded jobs being "posted" there by the master. Multiplexed in to the sequence of shared memory accesses initiated by master module 401, 402, 403, or 404 to post jobs there will be memory accesses by slaves which pop jobs off from the other stack. Once master module 401, 402, 403, or 404 has decoded and posted as jobs a complete video frame, and the slaves have entirely emptied the other job bank, the roles of the two banks flip, and the process begins again. Then, via block read operations, display controller 426 reads the data available in the dual port memory 504 for each module 401-404 over bus 425, and the reconstructed image information is placed into frame buffer 427.
The amount of computation needed to decode the original compressed file into "jobs" for the slaves is quite small when compared with the amount of subsequent calculation required to then process these jobs into completed areas of the final image. With four or fewer processors, the master module will almost always complete its decoding of the present frame before the slaves have emptied the other job banks. The master then reverts to the function of a slave, and joins the other slaves in finishing the remaining jobs. The result is that all available modules can be continuously employed until the frame is complete. A slave process will only fall idle under the circumstances that:
1. its last job is finished;
2. there are no remaining jobs for the frame; or
3. there are other slaves which have yet to finish their last jobs.
Since one job typically represents only 1/60 of the decode of the entire frame, it can be seen that decompression will be accomplished within the required time to be available to display controller 426, or frame buffer 430.
In each cycle through its main event loop, host 410 makes a circuit through all modules 401-404 and reads status registers. If a module has posted a task for host 410 (such as a request to input a frame of compressed data), host 410 takes the specified action, and continues.
The 4K longwords of dual port RAM 504 which is shared between host 410 and each module 401-404 is (logically) divided into two banks. This allows DSP 501 of each module 401-404 to be filling one bank in parallel with host 410 unloading the other into a frame buffer coupled to host 410. Alternatively, data may become available to frame buffer 430 which is accessed directly by modules 410-404. The rolls of the two banks can then be flipped when both processes are finished.
The preferred embodiment also provides a means for using reverse playback and still frame imagery using the CCITT H.261 compression standard. It will be appreciated by one skilled in the art, however, that this technique may be applied to other types of video compression such as MPEG or JPEG. As discussed previously, under CCITT H.261, forward motion video is made possible by the establishment of certain key or "intra" frames which establish the beginning of a scene. The frames following the "intra" frame and before the next "intra" frame, are known as "inter" frames which contain movement information for portions of the image. In other words, an inter frame only contains information for parts of the frame that has moved and that information is added to the intra frame information contained in frame buffer 430. However, because the "key" or "intra" frames only provide establishing information for the beginning of a scene, reverse playback is impossible, and an attempt to play an encoded image in reverse results in severe image distortions. The technique used for providing reverse playback and still frame in the preferred embodiment is shown and discussed with reference to FIG. 7a.
As shown in FIG. 7a, a particular series of frames, such as 700, may contain, as in the prior art discussed with reference to FIG. 3, several "forward facing" keyframes or "intra" frames 701, 702, and 703. Each of these keyframes provides complete scene information for the image at those particular points in time. In addition, there are several "inter" frames between the "intra" frames such as 710, 711, 712, and 713. Each of the "inter" frames provides motion information for image data which has moved since the last "intra" frame. The inter frames are added to the intra frame data contained in the frame buffer. For instance, 710 and 711 contain information that has changed since intra frame 701. Also, 712 and 713 contain motion information for parts of the image that has changed since intra frame 702. In the preferred embodiment, a series of images 700, however, also contains two extra frames 720 and 721, 720 and 721 are "additional" keyframes which have been added to images 700 to provide the reverse playback and still frame features. These additional keyframes establish a complete scene information in the reverse direction. In other words, 721 will establish complete scene information for the time, in a reverse direction, just prior to frames 713 and 712. While playing in a reverse direction, 721 will set the complete scene information of the image, and 713 will contain information which can be subtracted from keyframe 721, 712, in turn, contains information which has changed since inter frame 713, and can be subtracted from the image in the frame buffer. The same is true for backward facing keyframe 720 and its corresponding inter frames 711 and 710.
One distinguishing feature to note is that backward-facing keyframes 720 and 721 are ignored while playback is done in the forward direction. In other words, the additional keyframes are present in the sequence of compressed images, however, when played in the forward direction, keyframes 720 and 721 are skipped. This is because only intra frames 701, 702, and 703 are valid in the forward direction. Conversely, forward facing keyframes 701, 702, and 703 are skipped when the images are displayed in the reverse direction. Pointers are present in each of the frames 700 to point to the next frame in the sequence depending on the direction of play. In brief, pointers in the forward direction skip backward keyframes 720 and 721, and pointers in the reverse direction skip forward keyframes 701, 702, and 703. This is discussed in more detail below. In the reverse direction, only keyframes 721 and 720 are used to establish scene information from which inter frames are subtracted.
The enhanced CCITT encoder provides an additional keyframe where every forward facing keyframe appears, but this additional keyframe contains information which is only used for reverse play. The addition of the extra keyframe consumes approximately five percent more overhead in the computing of the additional keyframes. This is not considered significant, in light of the advantages provided by the reverse playback feature
A more detailed representation of a datum for a frame used in the preferred embodiment is shown in FIG. 7b. FIG. 7b shows "inter" frame 712 for the purposes of discussion, however, the discussion is equally applicable to the forward facing keyframes (intra frames) such as 701, 702, and 703 shown in FIG. 7a, and the backward facing keyframes such as 720 and 721 shown in FIG. 7a. As shown in FIG. 7b, a datum such as 712 contains three fields 751, 752, and 753 which provide information about the data contained within the frame. The first field, 751, is eight bits in length and known as a "frame ID" field. It contains a value indicating the type of frame. If the frame is an intra or forward-facing keyframe such as 701, 702, or 703, the frame ID field 751 contains zero. If, however, the frame is an inter frame such as 710, 711, 712, or 713 shown in FIG. 7a, then the frame ID contains one. Frame ID field 751 will contain two if the frame is a backward-facing keyframe such as 720 or 721 as provided by the preferred embodiment. The values 3 through 255 are currently undefined, therefore, the decoder of the preferred embodiment will ignore frames containing a frame ID field 751 with a value between 3 and 255, inclusive.
The next fields in the enhanced CCITT frame such as 712 shown in FIG. 7b, are the forward pointer 752 and the backward pointer 753. These fields merely provide linkage information for forward and reverse play. Reverse keyframes will be skipped for field 752 and forward keyframes (intra frames) will be skipped using field 753. In an intra frame such as 712 shown in FIG. 7a, the forward pointer 752 will point to frame 713, and the backward pointer will point to backward facing keyframe 720 as shown in FIG. 7a. Backward pointer 753 will point to other intra frames, such as 711 pointing to frame 710 shown in FIG. 7a, another inter frame precedes it in time. The remaining field in the datum such as 712 in FIG. 7b is variable length data field 754. This contains the appropriate variable length coding data for the frame. In the case of an intra frame such as 712 shown in FIG. 7a, the variable length data contains difference information from the previous frame such as 702 shown in FIG. 7a. For intra frames such as 701, 702, or 703, or backward facing keyframes such as 721 or 720, complete scene information is contained within the variable length data field 754.
The enhanced CCITT decoder is shown and discussed with reference to FIG. 8a. FIG. 8a shown a standard CCITT decoder with additional functional blocks added. Where the unmodified CCITT encoder comprised a motion estimation and intra-inter function blocks, the enhanced CCITT decoder 800 contains a frame difference block 814, a scene detector block 815, and a intra/inter/still/add block 816. Even though motion compensation is desirable because it removes more redundancy than frame differencing, it is very expensive in computing overhead. It is easier to implement the reverse playback feature using frame differencing. Scene detector block 815 automatically detects the difference in the energy off chrominance between successive frames. Also, block 815 detects scene changes and whether still images are present in the sequence of video images. Upon a scene change, key (intra) frames are added to the sequence to improve quality Block 815 decides whether the intra, inter, still, or "additional" frame (reverse keyframe) mode should be invoked. The additional frame mode is added to provide the necessary keyframe for reverse playback as discussed with reference to FIG. 8a. The frame difference block 814 takes the difference of consecutive frames rather than motion compensation to enable the reverse playback. Because there is no motion compensation as is provided in the forward direction, the quality of the image during compression and decompression is slightly degraded, however, this is acceptable considering the added features of reverse play and still frame along with the performance of CCITT H.261, which is adequate for many applications.
A more detailed representation of scene change detector 815 is shown in FIG. 8b. Scene change detector 815 takes difference information 852 received from frame difference block 814 on FIG. 8a and determines whether the difference from the previous frame is sufficient to warrant computation of an entirely new intra frame by block 816. This is determined by function 851, by computing the U and V (chrominance) energy contained within information 852 received from frame difference block 814. In an alternative embodiment, scene change detection may be keyed on luminance only, or both luminance and chrominance energy. Once the U and V information has been determined, that information is fed into a threshold block 850 which determines whether the U and V signals difference information has reached a predefined threshold. If the signal has reached this threshold, a signal is sent to block 816 shown in FIG. 8a to indicate that an entirely new intra or key frame must be computed for the input image to preserve the quality of the sequence. This indicates that the difference between the previous frame and the current frame is so great that entire scene information should be generated (an intra frame) instead of scene difference information contained in an inter frame. Therefore, the quality of the image and thus the sequence of moving images may be maintained. This information, which is sent to block 816 shown in FIG. 8a is output as information 853 shown in FIG. 8b.
The enhanced CCITT decoder which is used in the preferred embodiment is shown in FIG. 9. 900 as shown in FIG. 9 contains all the functions of the standard CCITT decoder, except that an additional block "play control" 908 is present to facilitate backwards and forwards play. Also, motion compensation block 206 has been replaced by frame differencing block 906 which performs the frame differencing for the uncompressed data depending on whether forward or reverse play is taking place. If forward play is taking place, then frame differencing block 906 merely adds inter frame data to the current data residing in the frame buffer. If reverse play is taking place, then frame differencing block 906 subtracts inter frame data from the current data residing in the frame buffer. Frame differencing block 906 and demultiplexing block 901 are controlled by play control block 908 which indicates to the decoder 900 whether forward or reverse play is taking place. Play control 908 is controlled by an external signal forward/backward play control signal 911 which is activated by user interaction on an input device such as a keyboard, mouse, or other input device.
Thus, an invention for improved video decompression has been described. Although this invention has been described particularly with reference to a preferred embodiment as set forth in FIGS. 1-9, it will be apparent to one skilled in the art that the present invention has utility far exceeding that disclosed in the figures. It is contemplated that many changes and modifications may be made, by one of ordinary skill in the art, without departing from the spirit and the scope of the invention as disclosed above. ##SPC1##
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4174514 *||26 Jun 1978||13 Nov 1979||Environmental Research Institute Of Michigan||Parallel partitioned serial neighborhood processors|
|US4484349 *||11 Mar 1982||20 Nov 1984||Environmental Research Institute Of Michigan||Parallel pipeline image processor|
|US4665556 *||9 Feb 1984||12 May 1987||Hitachi, Ltd.||Image signal processor|
|US4684997 *||25 Mar 1986||4 Aug 1987||Francoise Romeo||Machine for the reading, processing and compression of documents|
|US5070531 *||5 Nov 1990||3 Dec 1991||Oce-Nederland B.V.||Method of and means for processing image data|
|1||*||A. N. Netravali and B. G. Haskell, Digital Pictures: Representation and Compression, Plenum Publishing Corp., 1988, pp. 31 34, 389 392, and 415 416.|
|2||A. N. Netravali and B. G. Haskell, Digital Pictures: Representation and Compression, Plenum Publishing Corp., 1988, pp. 31-34, 389-392, and 415-416.|
|3||A. N. Netravali and J. O. Limb, "Picture Coding: A Review," Proceedings of the IEEE, vol. 68, No. 3, Mar. 1980, pp. 366-406.|
|4||*||A. N. Netravali and J. O. Limb, Picture Coding: A Review, Proceedings of the IEEE, vol. 68, No. 3, Mar. 1980, pp. 366 406.|
|5||C. Chen, "Adaptive Transform Coding Via Quadtree-Based Variable Blocksize DCT," ICASSP '89, IEEE, May 23-26, 1989, pp. 1854-1857.|
|6||*||C. Chen, Adaptive Transform Coding Via Quadtree Based Variable Blocksize DCT, ICASSP 89, IEEE, May 23 26, 1989, pp. 1854 1857.|
|7||E. J. Delp, O. R. Mitchell, "Image Compression Using Block Truncation Coding," IEEE Transactions on Communications, vol. COM-27, No. 9, Sep. 1979, pp. 1335-1342.|
|8||*||E. J. Delp, O. R. Mitchell, Image Compression Using Block Truncation Coding, IEEE Transactions on Communications, vol. COM 27, No. 9, Sep. 1979, pp. 1335 1342.|
|9||G. Campbell, T. A. Defanti, et al., "Two Bit/Pixel Full Color Encoding," Proceedings of SIGGRAPH '86, vol. 20, No. 4, Aug. 18-22, 1986, pp. 215-223.|
|10||*||G. Campbell, T. A. Defanti, et al., Two Bit/Pixel Full Color Encoding, Proceedings of SIGGRAPH 86, vol. 20, No. 4, Aug. 18 22, 1986, pp. 215 223.|
|11||International Organisation for Standardization, MPEG Video Simulation Model Three (SM3), Draft #1, A Koster, PTT Research, 1990.|
|12||*||International Organisation for Standardization, MPEG Video Simulation Model Three (SM3), Draft 1, A Koster, PTT Research, 1990.|
|13||*||J. B. O Neal, Jr., Predictive Quantizing Systems for the Transmission of Television Signals, Bell System Technical Journal, May/Jun. 1966, pp. 689 721.|
|14||J. B. O'Neal, Jr., "Predictive Quantizing Systems for the Transmission of Television Signals," Bell System Technical Journal, May/Jun. 1966, pp. 689-721.|
|15||*||N. S. Jayant and P. Noll, Digital Coding of Waveforms, Prentice Hall Signal Processing Series, p. 58.|
|16||N. S. Jayant and P. Noll, Digital Coding of Waveforms, Prentice-Hall Signal Processing Series, p. 58.|
|17||R. Plompen, Y. Hatori, et al., "Motion Video Coding in CCITT SG VS--The Video Source Coding," Globecom '88, Nov. 28-Dec. 1, 1988, pp. 0997-1004.|
|18||*||R. Plompen, Y. Hatori, et al., Motion Video Coding in CCITT SG VS The Video Source Coding, Globecom 88, Nov. 28 Dec. 1, 1988, pp. 0997 1004.|
|19||*||Specialists Group on Coding for Visual Telephony, Draft Revision of Recommendation H.261, Document No. 584, Nov. 10, 1989.|
|20||T. Murakami, K. Asai, et al., "Scene Adaptive Vector Quantization for Image Coding," Globecom '88, Nov. 28-Dec. 1, 1988, pp. 1068∝1072.|
|21||*||T. Murakami, K. Asai, et al., Scene Adaptive Vector Quantization for Image Coding, Globecom 88, Nov. 28 Dec. 1, 1988, pp. 1068 1072.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US5299019 *||18 Dec 1992||29 Mar 1994||Samsung Electronics Co., Ltd.||Image signal band compressing system for digital video tape recorder|
|US5467410 *||20 Mar 1992||14 Nov 1995||Xerox Corporation||Identification of a blank page in an image processing system|
|US5479527 *||8 Dec 1993||26 Dec 1995||Industrial Technology Research Inst.||Variable length coding system|
|US5502439 *||16 May 1994||26 Mar 1996||The United States Of America As Represented By The United States Department Of Energy||Method for compression of binary data|
|US5515480 *||15 Jun 1994||7 May 1996||Dp-Tek, Inc.||System and method for enhancing graphic features produced by marking engines|
|US5523847 *||9 Oct 1992||4 Jun 1996||International Business Machines Corporation||Digital image processor for color image compression|
|US5526025 *||19 Apr 1993||11 Jun 1996||Chips And Technolgies, Inc.||Method and apparatus for performing run length tagging for increased bandwidth in dynamic data repetitive memory systems|
|US5539664 *||20 Jun 1994||23 Jul 1996||Intel Corporation||Process, apparatus, and system for two-dimensional caching to perform motion estimation in video processing|
|US5553164 *||7 Oct 1994||3 Sep 1996||Canon Kabushiki Kaisha||Method for compressing and extending an image by transforming orthogonally and encoding the image|
|US5577191 *||17 Feb 1994||19 Nov 1996||Minerva Systems, Inc.||System and method for digital video editing and publishing, using intraframe-only video data in intermediate steps|
|US5579051 *||14 Dec 1994||26 Nov 1996||Mitsubishi Denki Kabushiki Kaisha||Method and apparatus for coding an input signal based on characteristics of the input signal|
|US5603012 *||7 Mar 1995||11 Feb 1997||Discovision Associates||Start code detector|
|US5617135 *||2 Sep 1994||1 Apr 1997||Hitachi, Ltd.||Multi-point visual communication system|
|US5623690 *||16 Jul 1992||22 Apr 1997||Digital Equipment Corporation||Audio/video storage and retrieval for multimedia workstations by interleaving audio and video data in data file|
|US5625571 *||7 Mar 1995||29 Apr 1997||Discovision Associates||Prediction filter|
|US5644504 *||27 Mar 1995||1 Jul 1997||International Business Machines Corporation||Dynamically partitionable digital video encoder processor|
|US5659654 *||2 Sep 1994||19 Aug 1997||Sony Corporation||Apparatus for recording and/or reproducing a video signal|
|US5669009 *||23 Feb 1996||16 Sep 1997||Hughes Electronics||Signal processing array|
|US5677981 *||14 Jun 1995||14 Oct 1997||Matsushita Electric Industrial Co., Ltd.||Video signal recording apparatus which receives a digital progressive scan TV signal and switches the progressive signal frame by frame alternately|
|US5689313 *||7 Jun 1995||18 Nov 1997||Discovision Associates||Buffer management in an image formatter|
|US5691767 *||24 Apr 1996||25 Nov 1997||Sony Corporation||Video conferencing system with high resolution still image capability|
|US5699460 *||17 Jun 1993||16 Dec 1997||Array Microsystems||Image compression coprocessor with data flow control and multiple processing units|
|US5699544 *||7 Jun 1995||16 Dec 1997||Discovision Associates||Method and apparatus for using a fixed width word for addressing variable width data|
|US5703793 *||7 Jun 1995||30 Dec 1997||Discovision Associates||Video decompression|
|US5710835 *||14 Nov 1995||20 Jan 1998||The Regents Of The University Of California, Office Of Technology Transfer||Storage and retrieval of large digital images|
|US5724537 *||6 Mar 1997||3 Mar 1998||Discovision Associates||Interface for connecting a bus to a random access memory using a two wire link|
|US5740460 *||7 Jun 1995||14 Apr 1998||Discovision Associates||Arrangement for processing packetized data|
|US5748776 *||5 Jul 1996||5 May 1998||Sharp Kabushiki Kaisha||Feature-region extraction method and feature-region extraction circuit|
|US5761741 *||7 Jun 1995||2 Jun 1998||Discovision Associates||Technique for addressing a partial word and concurrently providing a substitution field|
|US5764288 *||6 Jan 1995||9 Jun 1998||Integrated Data Systems, Inc.||Analog processing element (APE) and related devices|
|US5764801 *||27 Dec 1994||9 Jun 1998||Hitachi, Ltd.||Decoding system and method of coded data by parallel processings|
|US5768445 *||13 Sep 1996||16 Jun 1998||Silicon Graphics, Inc.||Compression and decompression scheme performed on shared workstation memory by media coprocessor|
|US5768561||7 Mar 1995||16 Jun 1998||Discovision Associates||Tokens-based adaptive video processing arrangement|
|US5768629||7 Jun 1995||16 Jun 1998||Discovision Associates||Token-based adaptive video processing arrangement|
|US5774676 *||3 Oct 1995||30 Jun 1998||S3, Incorporated||Method and apparatus for decompression of MPEG compressed data in a computer system|
|US5778096 *||12 Jun 1995||7 Jul 1998||S3, Incorporated||Decompression of MPEG compressed data in a computer system|
|US5784631 *||7 Mar 1995||21 Jul 1998||Discovision Associates||Huffman decoder|
|US5790712 *||8 Aug 1997||4 Aug 1998||8×8, Inc.||Video compression/decompression processing and processors|
|US5798719 *||7 Jun 1995||25 Aug 1998||Discovision Associates||Parallel Huffman decoder|
|US5801973 *||7 Jun 1995||1 Sep 1998||Discovision Associates||Video decompression|
|US5805914||7 Jun 1995||8 Sep 1998||Discovision Associates||Data pipeline system and data encoding method|
|US5808629 *||6 Feb 1996||15 Sep 1998||Cirrus Logic, Inc.||Apparatus, systems and methods for controlling tearing during the display of data in multimedia data processing and display systems|
|US5809201 *||21 Jun 1995||15 Sep 1998||Mitsubishi Denki Kabushiki Kaisha||Specially formatted optical disk and method of playback|
|US5809270||25 Sep 1997||15 Sep 1998||Discovision Associates||Inverse quantizer|
|US5815600 *||22 Aug 1995||29 Sep 1998||Hitachi Denshi Kabushiki Kaisha||Image data signal compression/transmission method and image data signal compression/transmission system|
|US5818524 *||11 Mar 1996||6 Oct 1998||Nikon Corporation||Digital still camera having an image data compression function|
|US5818967 *||12 Jun 1995||6 Oct 1998||S3, Incorporated||Video decoder engine|
|US5821885 *||7 Jun 1995||13 Oct 1998||Discovision Associates||Video decompression|
|US5825422 *||27 Dec 1996||20 Oct 1998||Daewoo Electronics Co. Ltd.||Method and apparatus for encoding a video signal based on inter-block redundancies|
|US5828907||7 Jun 1995||27 Oct 1998||Discovision Associates||Token-based adaptive video processing arrangement|
|US5829007 *||7 Jun 1995||27 Oct 1998||Discovision Associates||Technique for implementing a swing buffer in a memory array|
|US5831666 *||5 Aug 1996||3 Nov 1998||Digital Equipment Corporation||Video data scaling for video teleconferencing workstations communicating by digital data network|
|US5835740||7 Jun 1995||10 Nov 1998||Discovision Associates||Data pipeline system and data encoding method|
|US5835792||7 Jun 1995||10 Nov 1998||Discovision Associates||Token-based adaptive video processing arrangement|
|US5838380 *||23 Dec 1994||17 Nov 1998||Cirrus Logic, Inc.||Memory controller for decoding a compressed/encoded video data frame|
|US5844605 *||12 May 1997||1 Dec 1998||Integrated Data System, Inc.||Analog processing element (APE) and related devices|
|US5861894||7 Jun 1995||19 Jan 1999||Discovision Associates||Buffer manager|
|US5870497 *||28 May 1992||9 Feb 1999||C-Cube Microsystems||Decoder for compressed video signals|
|US5878273||7 Jun 1995||2 Mar 1999||Discovision Associates||System for microprogrammable state machine in video parser disabling portion of processing stages responsive to sequence-- end token generating by token generator responsive to received data|
|US5881301||2 Oct 1997||9 Mar 1999||Discovision Associates||Inverse modeller|
|US5898794 *||18 Jun 1997||27 Apr 1999||Fujitsu Limited||Image compression method and image processing system|
|US5903261 *||20 Jun 1996||11 May 1999||Data Translation, Inc.||Computer based video system|
|US5903674 *||29 Sep 1997||11 May 1999||Nec Corporation||Picture coding apparatus|
|US5905534 *||12 Jul 1994||18 May 1999||Sony Corporation||Picture decoding and encoding method and apparatus for controlling processing speeds|
|US5907692||24 Feb 1997||25 May 1999||Discovision Associates||Data pipeline system and data encoding method|
|US5923375 *||13 Feb 1997||13 Jul 1999||Sgs-Thomson Microelectronics S.R.L.||Memory reduction in the MPEG-2 main profile main level decoder|
|US5930386 *||1 Feb 1995||27 Jul 1999||Canon Kabushiki Kaisha||Image processing apparatus and method|
|US5930515 *||30 Sep 1997||27 Jul 1999||Scientific-Atlanta, Inc.||Apparatus and method for upgrading a computer system operating system|
|US5933572 *||24 Apr 1997||3 Aug 1999||Sony Corporation||Apparatus for recording and/or reproducing a video signal|
|US5937096 *||28 Nov 1995||10 Aug 1999||Canon Kabushiki Kaisha||Motion image processing apparatus and method|
|US5956519 *||1 May 1997||21 Sep 1999||Discovision Associates||Picture end token in a system comprising a plurality of pipeline stages|
|US5956741 *||15 Oct 1997||21 Sep 1999||Discovision Associates||Interface for connecting a bus to a random access memory using a swing buffer and a buffer manager|
|US5963673 *||18 Dec 1996||5 Oct 1999||Sanyo Electric Co., Ltd.||Method and apparatus for adaptively selecting a coding mode for video encoding|
|US5978545 *||8 Oct 1997||2 Nov 1999||Matsushita Electric Industrial Co., Ltd.||Video recording apparatus which accepts four different HDTV formatted signals|
|US5978592||8 Oct 1997||2 Nov 1999||Discovision Associates||Video decompression and decoding system utilizing control and data tokens|
|US5984512 *||7 Jun 1995||16 Nov 1999||Discovision Associates||Method for storing video information|
|US5987215 *||8 Oct 1997||16 Nov 1999||Matsushita Electric Industrial Co., Ltd.||Video signal recording apparatus, video signal recording and reproduction apparatus, video signal coding device, and video signal transmission apparatus|
|US5995727||7 Oct 1997||30 Nov 1999||Discovision Associates||Video decompression|
|US6002441 *||28 Oct 1996||14 Dec 1999||National Semiconductor Corporation||Audio/video subprocessor method and structure|
|US6011870 *||18 Jul 1997||4 Jan 2000||Jeng; Fure-Ching||Multiple stage and low-complexity motion estimation for interframe video coding|
|US6011900 *||8 Oct 1997||4 Jan 2000||Matsushita Electric Industrial Co., Ltd.||Video signal recording apparatus, video signal recording and reproduction apparatus, video signal coding device, and video signal transmission apparatus|
|US6018354 *||7 Jun 1995||25 Jan 2000||Discovision Associates||Method for accessing banks of DRAM|
|US6018776||21 Oct 1997||25 Jan 2000||Discovision Associates||System for microprogrammable state machine in video parser clearing and resetting processing stages responsive to flush token generating by token generator responsive to received data|
|US6035126||7 Jun 1995||7 Mar 2000||Discovision Associates||Data pipeline system and data encoding method|
|US6035322 *||2 Jan 1997||7 Mar 2000||Kabushiki Kaisha Toshiba||Image processing apparatus|
|US6038380||31 Jul 1997||14 Mar 2000||Discovision Associates||Data pipeline system and data encoding method|
|US6047112 *||7 Mar 1995||4 Apr 2000||Discovision Associates||Technique for initiating processing of a data stream of encoded video information|
|US6067417||7 Oct 1997||23 May 2000||Discovision Associates||Picture start token|
|US6070002 *||13 Sep 1996||30 May 2000||Silicon Graphics, Inc.||System software for use in a graphics computer system having a shared system memory|
|US6078721 *||1 Oct 1997||20 Jun 2000||Sharp K.K.||Video storage type communication device for selectively providing coded frames for selectable reproduction speed and mode according to type of a terminal device|
|US6079009||24 Sep 1997||20 Jun 2000||Discovision Associates||Coding standard token in a system compromising a plurality of pipeline stages|
|US6094453 *||16 Jan 1997||25 Jul 2000||Digital Accelerator Corporation||Digital data compression with quad-tree coding of header file|
|US6104751 *||26 Oct 1994||15 Aug 2000||Sgs-Thomson Microelectronics S.A.||Apparatus and method for decompressing high definition pictures|
|US6108015 *||2 Nov 1995||22 Aug 2000||Cirrus Logic, Inc.||Circuits, systems and methods for interfacing processing circuitry with a memory|
|US6112017||11 Nov 1997||29 Aug 2000||Discovision Associates||Pipeline processing machine having a plurality of reconfigurable processing stages interconnected by a two-wire interface bus|
|US6122726||3 Dec 1997||19 Sep 2000||Discovision Associates||Data pipeline system and data encoding method|
|US6124882 *||9 Jan 1998||26 Sep 2000||8×8, Inc.||Videocommunicating apparatus and method therefor|
|US6167190 *||1 Jul 1998||26 Dec 2000||Mitsubishi Denki Kabushiki Kaisha||Specially formatted optical disk and method of playback|
|US6195024||11 Dec 1998||27 Feb 2001||Realtime Data, Llc||Content independent data compression method and system|
|US6205248 *||27 Jan 1994||20 Mar 2001||Ncr Corporation||Method and system for compressing data in a multi-channel image processing system|
|US6205286 *||26 Mar 1997||20 Mar 2001||Hitachi, Ltd.||Image pickup apparatus having image information compressing function|
|US6217234||7 Jun 1995||17 Apr 2001||Discovision Associates||Apparatus and method for processing data with an arithmetic unit|
|US6219754 *||19 Dec 1997||17 Apr 2001||Advanced Micro Devices Inc.||Processor with decompressed video bus|
|US6226031 *||22 Oct 1998||1 May 2001||Netergy Networks, Inc.||Video communication/monitoring apparatus and method therefor|
|US6226413 *||22 Apr 1998||1 May 2001||Telefonaktiebolaget Lm Ericsson||Method for motion estimation|
|US6240492 *||22 May 1998||29 May 2001||International Business Machines Corporation||Memory interface for functional unit of integrated system allowing access to dedicated memory and shared memory, and speculative generation of lookahead fetch requests|
|US6263422||7 Jun 1995||17 Jul 2001||Discovision Associates||Pipeline processing machine with interactive stages operable in response to tokens and system and methods relating thereto|
|US6298145 *||19 Jan 1999||2 Oct 2001||Hewlett-Packard Company||Extracting image frames suitable for printing and visual presentation from the compressed image data|
|US6309424||3 Nov 2000||30 Oct 2001||Realtime Data Llc||Content independent data compression method and system|
|US6310974 *||1 Sep 2000||30 Oct 2001||Sharewave, Inc.||Method and apparatus for digital data compression|
|US6310975 *||1 Sep 2000||30 Oct 2001||Sharewave, Inc.||Method and apparatus for digital data compression|
|US6317134||20 Aug 1997||13 Nov 2001||Silicon Graphics, Inc.||System software for use in a graphics computer system having a shared system memory and supporting DM Pbuffers and other constructs aliased as DM buffers|
|US6326999||17 Aug 1995||4 Dec 2001||Discovision Associates||Data rate conversion|
|US6330665||10 Dec 1997||11 Dec 2001||Discovision Associates||Video parser|
|US6330666||7 Oct 1997||11 Dec 2001||Discovision Associates||Multistandard video decoder and decompression system for processing encoded bit streams including start codes and methods relating thereto|
|US6332041||18 Mar 1998||18 Dec 2001||Sharp Kabushiki Kaisha||Feature-region extraction method and feature-region extraction circuit|
|US6334021||20 Nov 2000||25 Dec 2001||Mitsubishi Denki Kabushiki Kaisha||Specially formatted optical disk and method of playback|
|US6356665||9 Dec 1998||12 Mar 2002||Sharp Laboratories Of America, Inc.||Quad-tree embedded image compression and decompression method and apparatus|
|US6360054||30 Dec 1999||19 Mar 2002||Sharp K.K.||Video storage type communication device|
|US6363076 *||27 Jan 1998||26 Mar 2002||International Business Machines Corporation||Phantom buffer for interfacing between buses of differing speeds|
|US6427194 *||30 Mar 2000||30 Jul 2002||Stmicroelectronics, Inc.||Electronic system and method for display using a decoder and arbiter to selectively allow access to a shared memory|
|US6435737||7 Jun 1995||20 Aug 2002||Discovision Associates||Data pipeline system and data encoding method|
|US6441842 *||16 Jun 1998||27 Aug 2002||8×8, Inc.||Video compression/decompression processing and processors|
|US6493466 *||13 Apr 1999||10 Dec 2002||Hitachi, Ltd.||Image data compression or expansion method and apparatus, and image transmission system and monitoring system using the method and device|
|US6499086||29 Jan 2001||24 Dec 2002||Advanced Micro Devices Inc.||Processor with decompressed video bus|
|US6526098||8 Oct 1997||25 Feb 2003||Matsushita Electric Industrial Co., Ltd.||High efficiency coding device and high efficiency coding method for image data|
|US6584222||19 Sep 2001||24 Jun 2003||Sharp Kabushiki Kaisha||Feature-region extraction method and feature-region extraction circuit|
|US6608938 *||4 Mar 2002||19 Aug 2003||Hitachi, Ltd.||Image data compression or expansion method and apparatus, and image transmission system and monitoring system using the method and device|
|US6624761||29 Oct 2001||23 Sep 2003||Realtime Data, Llc||Content independent data compression method and system|
|US6625309||29 Sep 1999||23 Sep 2003||Seiko Epson Corporation||Image partitioning to avoid overlap transmission|
|US6665318 *||13 May 1999||16 Dec 2003||Hitachi, Ltd.||Stream decoder|
|US6678331 *||2 Nov 2000||13 Jan 2004||Stmicrolectronics S.A.||MPEG decoder using a shared memory|
|US6678389||29 Dec 1998||13 Jan 2004||Kent Ridge Digital Labs||Method and apparatus for embedding digital information in digital multimedia data|
|US6697930||7 Feb 2001||24 Feb 2004||Discovision Associates||Multistandard video decoder and decompression method for processing encoded bit streams according to respective different standards|
|US6721455 *||8 May 1998||13 Apr 2004||Apple Computer, Inc.||Method and apparatus for icon compression and decompression|
|US6771884||3 May 2001||3 Aug 2004||Mitsubishi Denki Kabushiki Kaisha||Specially formatted optical disk and method of playback|
|US6799246||16 Dec 1997||28 Sep 2004||Discovision Associates||Memory interface for reading/writing data from/to a memory|
|US6826353||22 Apr 2003||30 Nov 2004||Mitsubishi Denki Kabushiki Kaisha||Specially formatted optical disk and method of playback|
|US6836564 *||6 Apr 2001||28 Dec 2004||Denso Corporation||Image data compressing method and apparatus which compress image data separately by modifying color|
|US6842578||8 May 2003||11 Jan 2005||Mitsubishi Denki Kabushiki Kaisha||Specially formatted optical disk and method of playback|
|US6847365 *||3 Jan 2000||25 Jan 2005||Genesis Microchip Inc.||Systems and methods for efficient processing of multimedia data|
|US6892296||1 Feb 2001||10 May 2005||Discovision Associates||Multistandard video decoder and decompression system for processing encoded bit streams including a standard-independent stage and methods relating thereto|
|US6910125||6 Feb 2001||21 Jun 2005||Discovision Associates||Multistandard video decoder and decompression system for processing encoded bit streams including a decoder with token generator and methods relating thereto|
|US6950930||5 Feb 2001||27 Sep 2005||Discovision Associates||Multistandard video decoder and decompression system for processing encoded bit streams including pipeline processing and methods relating thereto|
|US7000092 *||12 Dec 2002||14 Feb 2006||Lsi Logic Corporation||Heterogeneous multi-processor reference design|
|US7015868||12 Oct 2004||21 Mar 2006||Fractus, S.A.||Multilevel Antennae|
|US7058228||25 Jul 2003||6 Jun 2006||Hitachi, Ltd.||Image data compression or expansion method and apparatus, and image transmission system and monitoring system using the method and device|
|US7095783||12 Oct 2000||22 Aug 2006||Discovision Associates||Multistandard video decoder and decompression system for processing encoded bit streams including start codes and methods relating thereto|
|US7123208||8 Apr 2005||17 Oct 2006||Fractus, S.A.||Multilevel antennae|
|US7130913||28 Jul 2003||31 Oct 2006||Realtime Data Llc||System and methods for accelerated data storage and retrieval|
|US7149811||30 Jan 2001||12 Dec 2006||Discovision Associates||Multistandard video decoder and decompression system for processing encoded bit streams including a reconfigurable processing stage and methods relating thereto|
|US7158681||10 Feb 2003||2 Jan 2007||Cirrus Logic, Inc.||Feedback scheme for video compression system|
|US7161506||22 Sep 2003||9 Jan 2007||Realtime Data Llc||Systems and methods for data compression such as content dependent data compression|
|US7181125||9 Jan 2002||20 Feb 2007||Sharp K.K.||Video storage type communication device|
|US7181608||2 Feb 2001||20 Feb 2007||Realtime Data Llc||Systems and methods for accelerated loading of operating systems and application programs|
|US7227589||22 Dec 1999||5 Jun 2007||Intel Corporation||Method and apparatus for video decoding on a multiprocessor system|
|US7228064 *||2 Aug 2002||5 Jun 2007||Matsushita Electric Industrial Co., Ltd.||Image decoding apparatus, recording medium which computer can read from, and program which computer can read|
|US7230986||10 Oct 2001||12 Jun 2007||Discovision Associates||Multistandard video decoder and decompression system for processing encoded bit streams including a video formatter and methods relating thereto|
|US7321368||19 Jun 2002||22 Jan 2008||Stmicroelectronics, Inc.||Electronic system and method for display using a decoder and arbiter to selectively allow access to a shared memory|
|US7321937||8 Apr 2006||22 Jan 2008||Realtime Data Llc||System and methods for accelerated data storage and retrieval|
|US7352300||8 Jan 2007||1 Apr 2008||Realtime Data Llc||Data compression systems and methods|
|US7358867||8 Apr 2006||15 Apr 2008||Realtime Data Llc||Content independent data compression method and system|
|US7376184 *||4 Nov 2002||20 May 2008||Mitsubishi Denki Kabushiki Kaisha||High-efficiency encoder and video information recording/reproducing apparatus|
|US7376772||8 Apr 2006||20 May 2008||Realtime Data Llc||Data storewidth accelerator|
|US7378992||8 Apr 2006||27 May 2008||Realtime Data Llc||Content independent data compression method and system|
|US7386046||13 Feb 2002||10 Jun 2008||Realtime Data Llc||Bandwidth sensitive data compression and decompression|
|US7394432||17 Oct 2006||1 Jul 2008||Fractus, S.A.||Multilevel antenna|
|US7395345||8 Apr 2006||1 Jul 2008||Realtime Data Llc||System and methods for accelerated data storage and retrieval|
|US7397431||12 Jul 2005||8 Jul 2008||Fractus, S.A.||Multilevel antennae|
|US7400274||13 Mar 2007||15 Jul 2008||Realtime Data Llc||System and method for data feed acceleration and encryption|
|US7415530||26 Oct 2006||19 Aug 2008||Realtime Data Llc||System and methods for accelerated data storage and retrieval|
|US7417568||7 May 2003||26 Aug 2008||Realtime Data Llc||System and method for data feed acceleration and encryption|
|US7457518||29 Nov 2004||25 Nov 2008||Mitsubishi Denki Kabushiki Kaisha||Specially formatted optical disk and method of playback|
|US7505007||17 Oct 2006||17 Mar 2009||Fractus, S.A.||Multi-level antennae|
|US7526124 *||27 Feb 2007||28 Apr 2009||Intel Corporation||Match MSB digital image compression|
|US7528782||20 Jul 2007||5 May 2009||Fractus, S.A.||Multilevel antennae|
|US7542045||13 Dec 2007||2 Jun 2009||Stmicroelectronics, Inc.||Electronic system and method for display using a decoder and arbiter to selectively allow access to a shared memory|
|US7593623||14 Dec 2005||22 Sep 2009||Mitsubishi Denki Kabushiki Kaisha||Specially formatted optical disk and method of playback|
|US7669130||15 Apr 2005||23 Feb 2010||Apple Inc.||Dynamic real-time playback|
|US7710426||25 Apr 2005||4 May 2010||Apple Inc.||Buffer requirements reconciliation|
|US7711938||26 Jan 2001||4 May 2010||Adrian P Wise||Multistandard video decoder and decompression system for processing encoded bit streams including start code detection and methods relating thereto|
|US7714747||8 Jan 2007||11 May 2010||Realtime Data Llc||Data compression systems and methods|
|US7777651||2 Jun 2008||17 Aug 2010||Realtime Data Llc||System and method for data feed acceleration and encryption|
|US7777753||15 Apr 2009||17 Aug 2010||Stmicroelectronics, Inc.||Electronic system and method for selectively allowing access to a shared memory|
|US7783781||5 Oct 2005||24 Aug 2010||F5 Networks, Inc.||Adaptive compression|
|US7792194||10 Apr 2003||7 Sep 2010||Lefan Zhong||MPEG artifacts post-processed filtering architecture|
|US7873065||1 Feb 2006||18 Jan 2011||F5 Networks, Inc.||Selectively enabling network packet concatenation based on metrics|
|US7882084||25 Jan 2006||1 Feb 2011||F5 Networks, Inc.||Compression of data transmitted over a network|
|US7898548||16 Aug 2010||1 Mar 2011||Stmicroelectronics, Inc.||Electronic system and method for selectively allowing access to a shared memory|
|US7912349||25 Apr 2005||22 Mar 2011||Apple Inc.||Validating frame dependency information|
|US7983495 *||16 Aug 2007||19 Jul 2011||Fujitsu Semiconductor Limited||Image processing device and method|
|US8009111||10 Mar 2009||30 Aug 2011||Fractus, S.A.||Multilevel antennae|
|US8010668||29 Dec 2010||30 Aug 2011||F5 Networks, Inc.||Selective compression for network connections|
|US8024483||1 Oct 2004||20 Sep 2011||F5 Networks, Inc.||Selective compression for network connections|
|US8054315||27 Jan 2011||8 Nov 2011||Stmicroelectronics, Inc.||Electronic system and method for selectively allowing access to a shared memory|
|US8054879||8 Jan 2010||8 Nov 2011||Realtime Data Llc||Bandwidth sensitive data compression and decompression|
|US8073047||19 May 2008||6 Dec 2011||Realtime Data, Llc||Bandwidth sensitive data compression and decompression|
|US8090936||19 Oct 2006||3 Jan 2012||Realtime Data, Llc||Systems and methods for accelerated loading of operating systems and application programs|
|US8112619||19 Oct 2006||7 Feb 2012||Realtime Data Llc||Systems and methods for accelerated loading of operating systems and application programs|
|US8154462||28 Feb 2011||10 Apr 2012||Fractus, S.A.||Multilevel antennae|
|US8154463||9 Mar 2011||10 Apr 2012||Fractus, S.A.||Multilevel antennae|
|US8249367 *||30 Mar 2006||21 Aug 2012||Lg Electronics Inc.||Apparatus and method for encoding an image for a mobile telecommunication handset|
|US8275897||8 Apr 2006||25 Sep 2012||Realtime Data, Llc||System and methods for accelerated data storage and retrieval|
|US8275909||16 Mar 2006||25 Sep 2012||F5 Networks, Inc.||Adaptive compression|
|US8314808||21 Sep 2011||20 Nov 2012||Stmicroelectronics, Inc.||Electronic system and method for selectively allowing access to a shared memory|
|US8326984||18 Aug 2011||4 Dec 2012||F5 Networks, Inc.||Selective compression for network connections|
|US8330659||2 Mar 2012||11 Dec 2012||Fractus, S.A.||Multilevel antennae|
|US8350966 *||8 Jun 2009||8 Jan 2013||Broadcom Corporation||Method and system for motion compensated noise level detection and measurement|
|US8355587 *||11 Apr 2010||15 Jan 2013||Mediatek Inc.||Image processing apparatus capable of writing compressed data into frame buffer and reading buffered data from frame buffer alternately and related image processing method thereof|
|US8417833||29 Nov 2006||9 Apr 2013||F5 Networks, Inc.||Metacodec for optimizing network data compression based on comparison of write and read rates|
|US8437392||15 Apr 2005||7 May 2013||Apple Inc.||Selective reencoding for GOP conformity|
|US8477798||15 Dec 2010||2 Jul 2013||F5 Networks, Inc.||Selectively enabling network packet concatenation based on metrics|
|US8499100||21 Mar 2012||30 Jul 2013||F5 Networks, Inc.||Adaptive compression|
|US8502707||9 Feb 2010||6 Aug 2013||Realtime Data, Llc||Data compression systems and methods|
|US8504710||26 Oct 2006||6 Aug 2013||Realtime Data Llc||System and methods for accelerated data storage and retrieval|
|US8516113||25 Jul 2012||20 Aug 2013||F5 Networks, Inc.||Selective compression for network connections|
|US8516156||16 Jul 2010||20 Aug 2013||F5 Networks, Inc.||Adaptive compression|
|US8533308||5 Oct 2005||10 Sep 2013||F5 Networks, Inc.||Network traffic management through protocol-configurable transaction processing|
|US8553759||6 Jun 2011||8 Oct 2013||Realtime Data, Llc||Bandwidth sensitive data compression and decompression|
|US8559313||9 Sep 2011||15 Oct 2013||F5 Networks, Inc.||Selectively enabling packet concatenation based on a transaction boundary|
|US8565088||2 Mar 2006||22 Oct 2013||F5 Networks, Inc.||Selectively enabling packet concatenation based on a transaction boundary|
|US8611222||22 Aug 2012||17 Dec 2013||F5 Networks, Inc.||Selectively enabling packet concatenation based on a transaction boundary|
|US8643513||6 Jun 2011||4 Feb 2014||Realtime Data Llc||Data compression systems and methods|
|US8645834||5 Jan 2010||4 Feb 2014||Apple Inc.||Dynamic real-time playback|
|US8660182||9 Jun 2003||25 Feb 2014||Nvidia Corporation||MPEG motion estimation based on dual start points|
|US8660380||25 Aug 2006||25 Feb 2014||Nvidia Corporation||Method and system for performing two-dimensional transform on data value array with reduced power consumption|
|US8666166||30 Dec 2009||4 Mar 2014||Nvidia Corporation||Method and system for performing two-dimensional transform on data value array with reduced power consumption|
|US8666181||10 Dec 2008||4 Mar 2014||Nvidia Corporation||Adaptive multiple engine image motion detection system and method|
|US8681164||18 Oct 2012||25 Mar 2014||Stmicroelectronics, Inc.||Electronic system and method for selectively allowing access to a shared memory|
|US8692695||16 Aug 2010||8 Apr 2014||Realtime Data, Llc||Methods for encoding and decoding data|
|US8717203||24 Sep 2013||6 May 2014||Realtime Data, Llc||Data compression systems and methods|
|US8717204||24 Sep 2013||6 May 2014||Realtime Data Llc||Methods for encoding and decoding data|
|US8719438||5 May 2011||6 May 2014||Realtime Data Llc||System and methods for accelerated data storage and retrieval|
|US8723701||24 Sep 2013||13 May 2014||Realtime Data Llc||Methods for encoding and decoding data|
|US8724702||29 Mar 2006||13 May 2014||Nvidia Corporation||Methods and systems for motion estimation used in video coding|
|US8731071||15 Dec 2005||20 May 2014||Nvidia Corporation||System for performing finite input response (FIR) filtering in motion estimation|
|US8738103||21 Dec 2006||27 May 2014||Fractus, S.A.||Multiple-body-configuration multimedia and smartphone multifunction wireless devices|
|US8742958||24 Sep 2013||3 Jun 2014||Realtime Data Llc||Methods for encoding and decoding data|
|US8755675||11 Apr 2007||17 Jun 2014||Texas Instruments Incorporated||Flexible and efficient memory utilization for high bandwidth receivers, integrated circuits, systems, methods and processes of manufacture|
|US8756332||26 Oct 2006||17 Jun 2014||Realtime Data Llc||System and methods for accelerated data storage and retrieval|
|US8756482||25 May 2007||17 Jun 2014||Nvidia Corporation||Efficient encoding/decoding of a sequence of data frames|
|US8836713 *||3 Mar 2010||16 Sep 2014||Qualcomm Incorporated||Driving and synchronizing multiple display panels|
|US8867610||19 Dec 2013||21 Oct 2014||Realtime Data Llc||System and methods for video and audio data distribution|
|US8873625||18 Jul 2007||28 Oct 2014||Nvidia Corporation||Enhanced compression in representing non-frame-edge blocks of image frames|
|US8880862||27 May 2011||4 Nov 2014||Realtime Data, Llc||Systems and methods for accelerated loading of operating systems and application programs|
|US8929442||19 Dec 2013||6 Jan 2015||Realtime Data, Llc||System and methods for video and audio data distribution|
|US8933825||11 Apr 2014||13 Jan 2015||Realtime Data Llc||Data compression systems and methods|
|US8934535||20 Sep 2013||13 Jan 2015||Realtime Data Llc||Systems and methods for video and audio data storage and distribution|
|US8941541||2 Jan 2013||27 Jan 2015||Fractus, S.A.||Multilevel antennae|
|US8976069||2 Jan 2013||10 Mar 2015||Fractus, S.A.||Multilevel antennae|
|US8996996||29 Jan 2014||31 Mar 2015||Apple Inc.||Dynamic real-time playback|
|US9000985||2 Jan 2013||7 Apr 2015||Fractus, S.A.||Multilevel antennae|
|US9002806||8 Dec 2010||7 Apr 2015||F5 Networks, Inc.||Compression of data transmitted over a network|
|US9054421||2 Jan 2013||9 Jun 2015||Fractus, S.A.||Multilevel antennae|
|US9054728||24 Sep 2014||9 Jun 2015||Realtime Data, Llc||Data compression systems and methods|
|US9099773||7 Apr 2014||4 Aug 2015||Fractus, S.A.||Multiple-body-configuration multimedia and smartphone multifunction wireless devices|
|US9106606||16 Nov 2008||11 Aug 2015||F5 Networks, Inc.||Method, intermediate device and computer program code for maintaining persistency|
|US9116908||12 Jun 2014||25 Aug 2015||Realtime Data Llc||System and methods for accelerated data storage and retrieval|
|US9118927||13 Jun 2007||25 Aug 2015||Nvidia Corporation||Sub-pixel interpolation and its application in motion compensated encoding of a video signal|
|US9141992||23 Feb 2012||22 Sep 2015||Realtime Data Llc||Data feed acceleration|
|US9143546||3 Oct 2001||22 Sep 2015||Realtime Data Llc||System and method for data feed acceleration and encryption|
|US9210239||7 Mar 2013||8 Dec 2015||F5 Networks, Inc.||Metacodec for optimizing network data compression based on comparison of write and read rates|
|US9225479||13 Sep 2012||29 Dec 2015||F5 Networks, Inc.||Protocol-configurable transaction processing|
|US9236882||1 Jun 2015||12 Jan 2016||Realtime Data, Llc||Data compression systems and methods|
|US9240632||27 Jun 2013||19 Jan 2016||Fractus, S.A.||Multilevel antennae|
|US9330060||15 Apr 2004||3 May 2016||Nvidia Corporation||Method and device for encoding and decoding video image data|
|US9356824||29 Sep 2006||31 May 2016||F5 Networks, Inc.||Transparently cached network resources|
|US9362617||13 Aug 2015||7 Jun 2016||Fractus, S.A.||Multilevel antennae|
|US9606302||13 Mar 2014||28 Mar 2017||Commscope Technologies Llc||Ferrules for fiber optic connectors|
|US9614772||21 Nov 2003||4 Apr 2017||F5 Networks, Inc.||System and method for directing network traffic in tunneling applications|
|US9667751||14 Sep 2015||30 May 2017||Realtime Data, Llc||Data feed acceleration|
|US9761934||25 Apr 2016||12 Sep 2017||Fractus, S.A.||Multilevel antennae|
|US9762907||8 Jun 2015||12 Sep 2017||Realtime Adaptive Streaming, LLC||System and methods for video and audio data distribution|
|US9769477||6 Oct 2015||19 Sep 2017||Realtime Adaptive Streaming, LLC||Video data compression systems|
|US9792128||3 Nov 2014||17 Oct 2017||Realtime Data, Llc||System and method for electrical boot-device-reset signals|
|US20010036308 *||6 Apr 2001||1 Nov 2001||Osamu Katayama||Image data compressing method and apparatus which compress image data separately by modifying color|
|US20020035724 *||24 Jul 2001||21 Mar 2002||Wise Adrian Philip||Data rate conversion|
|US20020057739 *||18 Oct 2001||16 May 2002||Takumi Hasebe||Method and apparatus for encoding video|
|US20020057899 *||9 Jan 2002||16 May 2002||Sharp K.K.||Video storage type communication device|
|US20020066007 *||5 Feb 2001||30 May 2002||Wise Adrian P.||Multistandard video decoder and decompression system for processing encoded bit streams including pipeline processing and methods relating thereto|
|US20020069354 *||2 Feb 2001||6 Jun 2002||Fallon James J.||Systems and methods for accelerated loading of operating systems and application programs|
|US20020180743 *||19 Jun 2002||5 Dec 2002||Stmicroelectronics, Inc.|
|US20020191692 *||13 Feb 2002||19 Dec 2002||Realtime Data, Llc||Bandwidth sensitive data compression and decompression|
|US20030026487 *||2 Aug 2002||6 Feb 2003||Yoshiyuki Wada||Image decoding apparatus, recording medium which computer can read from, and program which computer can read|
|US20030133501 *||4 Nov 2002||17 Jul 2003||Mitsubishi Denki Kabushiki Kaisha||High-efficiency encoder and video information recording/reproducing apparatus|
|US20030182544 *||6 Feb 2001||25 Sep 2003||Wise Adrian P.||Multistandard video decoder and decompression system for processing encoded bit streams including a decoder with token generator and methods relating thereto|
|US20030191876 *||27 Nov 2002||9 Oct 2003||Fallon James J.||Data storewidth accelerator|
|US20030196078 *||8 Feb 2001||16 Oct 2003||Wise Adrian P.||Data pipeline system and data encoding method|
|US20040025000 *||26 Jan 2001||5 Feb 2004||Wise Adrian P.||Multistandard video decoder and decompression system for processing encoded bit streams including start code detection and methods relating thereto|
|US20040056783 *||22 Sep 2003||25 Mar 2004||Fallon James J.||Content independent data compression method and system|
|US20040073746 *||28 Jul 2003||15 Apr 2004||Fallon James J.||System and methods for accelerated data storage and retrieval|
|US20040117743 *||12 Dec 2002||17 Jun 2004||Judy Gehman||Heterogeneous multi-processor reference design|
|US20040156549 *||10 Feb 2003||12 Aug 2004||Cirrus Logic, Inc.||Feedback scheme for video compression system|
|US20040218672 *||11 Jul 2002||4 Nov 2004||Bourne David Ronald||Video transmission system video transmission unit and methods of encoding decoding video data|
|US20040221143 *||1 Feb 2001||4 Nov 2004||Wise Adrian P.||Multistandard video decoder and decompression system for processing encoded bit streams including a standard-independent stage and methods relating thereto|
|US20040233986 *||28 Jun 2004||25 Nov 2004||Amir Morad||Video encoding device|
|US20040240744 *||25 Jul 2003||2 Dec 2004||Toyota Honda|
|US20040247029 *||9 Jun 2003||9 Dec 2004||Lefan Zhong||MPEG motion estimation based on dual start points|
|US20040247034 *||10 Apr 2003||9 Dec 2004||Lefan Zhong||MPEG artifacts post-processed filtering architecture|
|US20050094974 *||29 Nov 2004||5 May 2005||Mitsubishi Denki Kabushiki Kaisha||Specially formatted optical disk and method of playback|
|US20050110688 *||12 Oct 2004||26 May 2005||Baliarda Carles P.||Multilevel antennae|
|US20050259009 *||8 Apr 2005||24 Nov 2005||Carles Puente Baliarda||Multilevel antennae|
|US20060015650 *||19 Sep 2005||19 Jan 2006||Fallon James J||System and methods for accelerated data storage and retrieval|
|US20060038888 *||16 Aug 2005||23 Feb 2006||Olympus Corporation||Image transmission apparatus|
|US20060098953 *||14 Dec 2005||11 May 2006||Masato Nagasawa||Specially formatted optical disk and method of playback|
|US20060184687 *||8 Apr 2006||17 Aug 2006||Fallon James J||System and methods for accelerated data storage and retrieval|
|US20060184696 *||8 Apr 2006||17 Aug 2006||Fallon James J||System and methods for accelerated data storage and retrieval|
|US20060233237 *||15 Apr 2005||19 Oct 2006||Apple Computer, Inc.||Single pass constrained constant bit-rate encoding|
|US20060233245 *||15 Apr 2005||19 Oct 2006||Chou Peter H||Selective reencoding for GOP conformity|
|US20060236245 *||15 Apr 2005||19 Oct 2006||Sachin Agarwal||Dynamic real-time playback|
|US20060239565 *||30 Mar 2006||26 Oct 2006||Lg Electronics Inc.||Apparatus and method for encoding an image for a mobile telecommunication handset|
|US20060274955 *||10 May 2006||7 Dec 2006||Toyota Honda|
|US20060290573 *||12 Jul 2005||28 Dec 2006||Carles Puente Baliarda||Multilevel antennae|
|US20070019875 *||21 Jul 2005||25 Jan 2007||Sung Chih-Ta S||Method of further compressing JPEG image|
|US20070067483 *||26 Oct 2006||22 Mar 2007||Realtime Data Llc||System and methods for accelerated data storage and retrieval|
|US20070147692 *||27 Feb 2007||28 Jun 2007||Dwyer Michael K||Match MSB digital image compression|
|US20070247936 *||11 Apr 2007||25 Oct 2007||Texas Instruments Incorporated||Flexible and efficient memory utilization for high bandwidth receivers, integrated circuits, systems, methods and processes of manufacture|
|US20080044088 *||16 Aug 2007||21 Feb 2008||Fujitsu Limited||Image processing device and method|
|US20080050036 *||25 Aug 2006||28 Feb 2008||Portalplayer, Inc.||Method and system for performing two-dimensional transform on data value array with reduced power consumption|
|US20080088637 *||13 Dec 2007||17 Apr 2008||Stmicroelectronics, Inc.|
|US20080291209 *||25 May 2007||27 Nov 2008||Nvidia Corporation||Encoding Multi-media Signals|
|US20080294962 *||25 May 2007||27 Nov 2008||Nvidia Corporation||Efficient Encoding/Decoding of a Sequence of Data Frames|
|US20080310509 *||13 Jun 2007||18 Dec 2008||Nvidia Corporation||Sub-pixel Interpolation and its Application in Motion Compensated Encoding of a Video Signal|
|US20090022219 *||18 Jul 2007||22 Jan 2009||Nvidia Corporation||Enhanced Compression In Representing Non-Frame-Edge Blocks Of Image Frames|
|US20090167625 *||10 Mar 2009||2 Jul 2009||Fractus, S.A.||Multilevel antennae|
|US20090201305 *||15 Apr 2009||13 Aug 2009||Stmicroelectronics, Inc.||Electronic system and method for selectively allowing access to a shared memory|
|US20100104008 *||30 Dec 2009||29 Apr 2010||Nvidia Corporation|
|US20100104023 *||15 Dec 2009||29 Apr 2010||Smith Ronald D||Compressing Video Frames|
|US20100142761 *||10 Dec 2008||10 Jun 2010||Nvidia Corporation||Adaptive multiple engine image motion detection system and method|
|US20100309211 *||16 Aug 2010||9 Dec 2010||Stmicroelectronics, Inc.||Electronic system and method for selectively allowing access to a shared memory|
|US20100309378 *||8 Jun 2009||9 Dec 2010||Sheng Zhong||Method And System For Motion Compensated Noise Level Detection And Measurement|
|US20110122946 *||27 Jan 2011||26 May 2011||Stmicroelectronics, Inc.||Electronic system and method for selectively allowing access to a shared memory|
|US20110163923 *||9 Mar 2011||7 Jul 2011||Fractus, S.A.||Multilevel antennae|
|US20110175777 *||28 Feb 2011||21 Jul 2011||Fractus, S.A.||Multilevel antennae|
|US20110216082 *||3 Mar 2010||8 Sep 2011||Qualcomm Incorporated||Driving and synchronizing multiple display panels|
|US20110249906 *||11 Apr 2010||13 Oct 2011||Chiuan-Shian Chen||Image processing apparatus capable of writing compressed data into frame buffer and reading buffered data from frame buffer alternately and related image processing method thereof|
|CN102215382A *||8 Mar 2011||12 Oct 2011||联发科技股份有限公司||Image processing apparatus and image processing method|
|CN104768021A *||22 Apr 2015||8 Jul 2015||四川正冠科技有限公司||Ultra-low time delay H.264 coding method and coder|
|EP0614317A2 *||22 Feb 1994||7 Sep 1994||Sony Corporation||Video signal decoding|
|EP0614317A3 *||22 Feb 1994||25 Jan 1995||Sony Corp||Video signal decoding.|
|EP0642274A2 *||5 Sep 1994||8 Mar 1995||Sony Corporation||Video signal recording/reproducing apparatus|
|EP0642274A3 *||5 Sep 1994||26 Apr 1995||Sony Corp||Video signal recording/reproducing apparatus.|
|EP0651579A1 *||25 Oct 1994||3 May 1995||Sgs-Thomson Microelectronics S.A.||High resolution image processing system|
|EP0651582A2 *||21 Oct 1994||3 May 1995||Philips Electronics N.V.||Device for transmitting television pictures and device for receiving said pictures|
|EP0651582A3 *||21 Oct 1994||19 Jul 1995||Philips Electronics Nv||Device for transmitting television pictures and device for receiving said pictures.|
|EP0658053A1 *||28 Jun 1994||14 Jun 1995||Sony Corporation||Apparatus for decoding time-varying image|
|EP0658053A4 *||28 Jun 1994||24 Apr 1996||Sony Corp||Apparatus for decoding time-varying image.|
|EP0660614A1 *||12 Jul 1994||28 Jun 1995||Sony Corporation||Method and apparatus for decoding image and method and apparatus for encoding image|
|EP0660614A4 *||12 Jul 1994||18 Mar 1998||Sony Corp||Method and apparatus for decoding image and method and apparatus for encoding image.|
|EP0688134A3 *||14 Jun 1995||14 Aug 1996||Matsushita Electric Ind Co Ltd||Video signal recording apparatus, video signal recording and reproduction apparatus, video signal coding device, and video signal transmission apparatus|
|EP0689355A3 *||23 Jun 1995||27 Nov 1996||Mitsubishi Electric Corp||Optical disk and method of playback|
|EP0701368A2||22 Aug 1995||13 Mar 1996||Discovision Associates||Data rate conversion|
|EP0720372A1 *||30 Dec 1994||3 Jul 1996||Daewoo Electronics Co., Ltd||Apparatus for parallel encoding/decoding of digital video signals|
|EP0720374A1 *||30 Dec 1994||3 Jul 1996||Daewoo Electronics Co., Ltd||Apparatus for parallel decoding of digital video signals|
|EP0891088A1 *||28 Feb 1995||13 Jan 1999||Discovision Associates||Pipeline decoding system|
|EP0896478A2 *||27 Jun 1994||10 Feb 1999||Kabushiki Kaisha Toshiba||Video decoder|
|EP0896478A3 *||27 Jun 1994||15 Sep 1999||Kabushiki Kaisha Toshiba||Video decoder|
|EP0940988A2 *||2 Mar 1999||8 Sep 1999||Sony Corporation||Digital television signal encoding and/or decoding|
|EP0940988A3 *||2 Mar 1999||9 Apr 2003||Sony Corporation||Digital television signal encoding and/or decoding|
|EP1282317A2 *||23 Jun 1995||5 Feb 2003||Mitsubishi Denki Kabushiki Kaisha||Optical disk and method of playback|
|EP1282317A3 *||23 Jun 1995||12 Feb 2003||Mitsubishi Denki Kabushiki Kaisha||Optical disk and method of playback|
|EP1315385A1 *||23 Jun 1995||28 May 2003||Mitsubishi Denki Kabushiki Kaisha||Optical disk and method of playback|
|EP1408701A1 *||23 Jun 1995||14 Apr 2004||Mitsubishi Denki Kabushiki Kaisha||Optical disk and method of playback|
|EP1411734A1 *||23 Jun 1995||21 Apr 2004||Mitsubishi Denki Kabushiki Kaisha||Optical disk and method of playback|
|EP1531631A2 *||23 Jun 1995||18 May 2005||Mitsubishi Denki Kabushiki Kaisha||Optical disk and method of playback|
|EP1531631A3 *||23 Jun 1995||25 May 2005||Mitsubishi Denki Kabushiki Kaisha||Optical disk and method of playback|
|EP1628486A1 *||17 Aug 2005||22 Feb 2006||Olympus Corporation||Image transmission apparatus|
|WO1996020567A1 *||20 Dec 1995||4 Jul 1996||Cirrus Logic, Inc.||Memory controller for decoding and displaying compressed video data|
|WO2001047284A1 *||20 Nov 2000||28 Jun 2001||Intel Corporation||Method and apparatus for video decoding on a multiprocessor system|
|U.S. Classification||382/166, 375/E07.211, 375/E07.256, 386/E09.015, 375/E07.088, 375/E07.093, 375/E07.103, 375/E07.277, 375/E07.148, 375/E07.224, 375/E07.094, 375/E07.275, 382/234, 375/E07.263|
|International Classification||H04N21/434, H04N21/236, H04N7/54, H04N11/00, H04N7/50, H04N7/36, H04N11/24, G06T9/00, H04N5/85, H04N9/804, H04N7/26, H04N5/783|
|Cooperative Classification||H04N19/30, H04N19/503, H04N19/107, H04N19/436, H04N19/423, H04N19/42, H04N19/51, H04N19/61, H04N21/236, H04N5/783, H04N7/54, H04N9/8047, H04N21/434, H04N5/85|
|European Classification||H04N21/236, H04N21/434, H04N7/26L2, H04N7/54, H04N7/50R, H04N7/26E, H04N7/36D, H04N7/50, H04N7/26L, H04N7/36C, H04N7/26A4C2, H04N9/804B3, H04N7/26L6|
|24 May 1991||AS||Assignment|
Owner name: APPLE COMPUTER, INC., A DE CORPORATION, CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNORS:NORMILE, JAMES O.;YEH, CHIA L.;WRIGHT, DANIEL W.;AND OTHERS;REEL/FRAME:005718/0315;SIGNING DATES FROM 19910517 TO 19910522
|5 Feb 1993||AS||Assignment|
Owner name: APPLE COMPUTER, INC., CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNORS:NORMILE, JAMES O.;YEH, CHIA LUNG;WRIGHT, DANIEL W.;AND OTHERS;REEL/FRAME:006402/0394;SIGNING DATES FROM 19930120 TO 19930201
|30 Sep 1996||FPAY||Fee payment|
Year of fee payment: 4
|17 Nov 2000||FPAY||Fee payment|
Year of fee payment: 8
|22 Sep 2004||FPAY||Fee payment|
Year of fee payment: 12
|11 May 2007||AS||Assignment|
Owner name: APPLE INC., CALIFORNIA
Free format text: CHANGE OF NAME;ASSIGNOR:APPLE COMPUTER, INC., A CALIFORNIA CORPORATION;REEL/FRAME:019304/0167
Effective date: 20070109